SDMS Document 
115572 
US. DEPARTWENTOF COMMERCE 
NaticxialOceanicarKJAbrioGphericAdrmisiration 
National Ooean Savioe 
OfoeofOoeanResajDesConsavaficnand/\ssessrnert 
IHazai±)us Materials Resporisearri/^ssessmertDMsion 
Coasial ResoucesCoorciTEfcn Branch 
Room3137-C 
26 Federal Plaza 
NewYork, New York 10278 
September?, 1995 
Sherrel Henry 
New Yorl^Caribbean Superfund Branch II 
Emergency and Remedial Response Division 
U.S. Environmental Protection Agency 
290 Broadway 
New York, NY 10007 
Dear Sherrel: 
Thank you for the opportunity to review the Draft Data Summary Report for the Virgin 
Island Chemical Site in St. Croix, U.S. Virgin Islands. The following comments are 
submitted by the National Oceanic and Atmospheric Administration (NOAA). The 
comments are broken into two sections which focus first on the Data Summary and then on 
the Ecological Assessment (Appendix L). 
Data Summary: 
The Draft Data Summary Report for the Virgin Island Chemical (VICHEM) Site provided 
results of the Remedial Investigation (RI) sampling required by the Final RI Work Plan. 
NOAA had the opportunity to review the Draft RI Work Plan in April 1994 and had 
expressed concern that the work plan only required soil and groundwater sampling onsite and 
did not require any sampling in the River Gut, the intermittent stream bordering the site to 
the northeast and southeast. NOAA felt that sampling the River Gut was essential for two 
reasons: 1) ail surface runoff from the site as well as groundwater discharge drains into the 
gut and 2) sediment sampling in the River Gut conducted as part of the Site Investigation (SI) 
effort for the site and summarized in the RI Work Plan documented the presence of several 
metals at elevated levels, especially zinc, cadmium, copper and nickel. The BTAG submitted 
comments noting similar concerns. Though we had several conversations about this issue 
following the submission of our comments, specifically pertaining to the issue of the 
presence of metals in onsite soils which NOAA had referred you to review the data in the 
Draft RI Work Plan, to NOAA's disappointment the Final RI Work Plan did not require any 
3 0 5 2 5 6 
4 
sampling in the River Gut. This is of great concern to NOAA as the River Gut drains into the 
Caribbean Sea approximately 1.4 kilometers downstream of the site. NOAA trust resources 
would be expected to utilize the Caribbean Sea near the mouth of the River Gut as well as 
areas closer to the site if mangrove or tidal habitats are present upstream from the' Caribbean 
Sea. 
The Draft Data Summary Report presents the results of the groundwater and onsite soil 
sampling completed to date. To evaluate groundwater quality, results were compared to 
Drinking Water Maximum Contaminant Levels (MCLs), RCRA Maximum Concentrations 
of Constituents for Groundwater Protection (RCRA MCCs) and RCRA Proposed Examples 
of Concentrations Meeting Criteria for Action Levels (RCRA ALs). According to the report, 
only five metals were detected at concentrations from filtered samples that exceeded one or 
more of the ARARs identified. These included antimony, beryllium, iron, manganese and 
selenium. No PCBs were detected in groundwater and the presence of pesticides were 
attributed to widespread use of them across the island. The report notes that the maximum 
concentrations for each of the substances detected in groundwater samples are presented with 
the ARARs on Table 5-3. No such table could be located in the report. Because several 
metals were detected above ARARs, as well as ethylbenzene, additional groundwater 
sampling is recommended. 
To evaluate soil quality, results of soil samples were compared to RCRA Examples of 
Concentrations Meeting Criteria for Action Levels (RCRA ALs). In addition, results of work 
completed by USGS which included sampling of 127 soil samples fi"om 1983 to 1988 were 
used to represent background metals concentrations for the island. The report notes that 
Table 5-2 presents a comparison of ranges of concentrations of metals in soil samples to the 
USGS results. No such table could be located in the report. The report noted that twenty 
TAL metals were detected in soil samples. Results for 13 of the metals detected were within 
or almost identical to ranges of concentrations reported by USGS. Three of the metals 
detected were not detected by the USGS sampling. These included aluminum, mercury and 
potassium. It was unlcear whether soil levels for these metals exceeded their respective 
RCRA ALs. Of the twenty metals reported, zinc was reported at concentrations higher than 
those reported by USGS. Zinc was detected in all samples collected at the site at 
concentrations ranging from 35.1 to 5790 mg/kg (the RCRA MDL for zinc being 200 
mg/kg). The report notes that on St. Croix, the soils are derived from volcaniclastics, which 
often contain heavy metals and that based on comparison with USGS data, the metals 
detected, with the exception of zinc, appear to represent background concentrations. The 
report further notes that elemental zinc was not reported on the available inventories of 
materials handled at the facility. Six pesticides and no PCBs were detected in the soil 
samples collected. The report concludes that the pesticides were detected sporadically and at 
305257 
f 
trace concentration and that based on widespread use of pesticides and the relatively low 
concentrations detected onsite, pesticides are not considered a concern. No additional 
activities relating to the soil are proposed. 
Previous sampling during the SI for this site documented elevated levels of pesticides, lead, 
copper and zinc present in soil from the Above-Ground Storage Tank Farm, elevated levels 
of lead and zinc in sediment from the Former Process Pit, elevated levels of zinc, cadmium, 
copper and nickel from River Gut sediment and elevated levels of zinc, lead, copper, nickel 
and cadmium from sediment collected from storm drains onsite. It appears that the levels for 
pesticides noted in the SI were erroneously as mg/kg rather than ug/kg however, the pattern 
of elevated levels of metals in onsite soil, storm drains and the River Gut strongly suggests 
that migration of site related contaminants via surface runoff has occurred. The more recent 
data collected during the RI effort present in this report and in the Ecological Assessment 
(EA) report suggests there is the potential for metals, especially zinc, to migrate offsite and 
possibly accumulate in the sediment in the River Gut where they may be bioavailable to 
aquatic receptors present in the River Gut or where they may be acting as a source of 
contamination further downstream to the Caribbean Sea. Contaminants discharged to the 
River Gut during the rainy season, September to December, would be available to aquatic 
receptors and during the dry season would be available to terrestrial receptors. During 
subsequent rainy seasons, contaminants bound to sediments/soils would potentially be 
carried ftirther downstream toward the Caribbean Sea. 
The report appears to have eliminated the need to further evaluate any risk associated with 
the presence of zinc and its' potential for migration offsite simply due to the fact that 
elemental zinc was not reported on the available inventories of materials handled by the 
facility. It is obvious from the sampling conducted at the site during the SI and the RI that 
this contaminant is site-related and that additional sampling of any migration pathways from 
the site (including storm drains and the River Gut) is required. By restricting the ecological 
investigation to onsite is ignoring the greatest potential pathway for contaminant migration to 
ecological receptors. 
Appendix L Phase I Baseline Ecological Assessment 
Three of the four potential sources of site-related contaminants proposed were later 
eliminated as threats to onsite ecological receptors. More information is needed about 
sources 2, 3, and 4 before eliminating them as potential threats. For example, it was 
determined that turning off the leaking valve of tank #5 (source #2) eliminated the source of 
contamination. However, no information was provided about the contents of tank #5, how 
long it had been leaking, the depth of contaminated soils, or whether an attempt was made to 
305258 
remove the residual contamination. Secondly, the elimination of contaminated soils (source 
#3) as a potential threat because of the limited distribution of contamination is inappropriate. 
In the event of localized hotspots of contamination, where contamination exceeds screening 
criteria, a source of exposure exists. Lastly, the elimination of the concrete settling tank 
(source #4) as a source of contaminants without more detailed description of the contents and 
condition of the tank, is premature. Therefore, from the information provided in the EA, 
sources 1, 2, 3, and 4 should all be included as potential sources of contamination. 
Of the three potential pathways for contaminant availability identified, the second one, 
transfer of contaminated soils or sediment via surface runoff, would more accurately be 
described as a source of contamination rather than a pathway of availability. Regardless, 
elimination of pathways #2 (runoff) and #3 (foodchain) was not appropriate. It is unclear 
how it was concluded that the distribution of contaminants by stormwater runoff was 
minimal. It was described on page 5 of the EA that a 4-hour rainstorm created stormwater 
flow "across the site" from 6 inches to 2 feet deep. Stormwater runoff of such magnitude 
would potentially transfer a significant amount of soils and sediments. Secondly, the 
foodchain as a pathway for availability of contaminants can not be eliminated without 
measuring tissue concentrations of potentially contaminated food sources and applying them 
to the appropriate foodchain models. Therefore, it is highly probably that surfacewater 
runoff could carry contaminants from their source to other areas on and offsite and that 
contaminants could be transferred through the foodchain. 
The three potential ecological receptors identified, ingestion of contaminated vegetation, 
ingestion and inhalation of contaminated soils, and ingestion of contaminated water, would 
more appropriately be included as additional pathways for availability of contaminants rather 
than potential ecological receptors. In the Environmental Resource Inventory (Attachment 
A) a few bird and mammal species present onsite were identified, but no information was 
provided regarding potential threats to these species or how they were related to ecological 
risks. Of particular concern is the herd of free roaming domestic goats "... observed 
regularly grazing [onsite] and drinlcing water from the open drainage channels around the 
process area." 
Regardless of the inaccurate categorizing of assessment terminology, the elirnination of all 
three potential receptors was premature. Itwas simply stated that vegetation growing in 
contaminated soils (receptor #1) is removed "occasionally", but it was not stated how 
frequent the removal occurs or how quickly the vegetation reinvades the removal areas. 
Without this information, it is impossible to determine if contaminated vegetation provides a 
source of contaminants to ecological receptors. Secondly, the incidental ingestion or 
inhalation of contaminated soils (receptor #2) can not be eliminated as a threat due to a 
3 0 5 2 5 8 A 
f 
limited distribution of contaminated soils. Localized hotspots of contamination may pose a 
threat to ecological receptors, particularly those with a limited home range. Lastly, the 
elimination of ingestion of contaminated water and vegetation from the process pit area 
(receptor #3) can not be justified without current sampling of the water and sediment in the 
drainage channels. It was stated in the EA that historic contaminants of the process pit had 
been "flushed" out of the system several times since a 1985 sampling event showed elevated 
concentrations of several metals. This argument supports sarhpling sedirnents in River Gut, 
the receptor for "flushed" contaminants. 
Conclusion 
From the information provided in the Data Summary report and the EA, the potential for 
offsite migration of contaminants, particularly zinc, is highly probable. Therefore, sediment 
sampling in the River Gut downstream to the Caribbean Sea is needed to determine the 
extent of contaminant migration. At that time, the need for a fiill scale off-site ecological risk 
assessment would be evaluated to determine the potential risks to ecological receptors 
downstream of the site. In addition, based on the information presented in the EA, it can not 
be concluded that on-site ecological risks are minimal. The onsite ecological assessment was 
incomplete and, therefore, inconclusive. 
NOAA would appreciate being kept informed as additional information becomes available at 
the VICHEM site. Should you have any questions regarding these comments, please feel 
free to contact either Jermifer Scott or me at (212) 637-3259. 
Sincerely, 
Diane E. Wehner 
NOAA Coastal Resource Coordinator 
cc: Carole Petersen, NYCSB 
Shari Stevens, BTAG 
3 0 5 2 5 9